SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 2306.00977 · ICLR · 2024

AGILE3D: Attention Guided Interactive Multi-object 3D Segmentation

Bastian Leibe, Konrad Schindler, Yuanwen Yue, Sabarinath Mahadevan, Jonas Schult, Francis Engelmann, Theodora Kontogianni

arXiv · PDF · Open in the Atlas

Code that ran

7 of 7 samples (functions and classes) lifted from this paper's linked repositories executed on a synthesized input in Syntology's jail (0 of them constructed an object rather than computing a result; 0 honoured the contract we drafted, 0 violated it and 6 have no contract verdict; 1 failure of our instrument; 0 no recorded run), measured 2026-11-22; repository links from the Papers with Code archive attributed under CC BY-SA, no affiliation; bodies inline only under a permissive licence as GitHub detects it. Not a reproduction of the paper's results.

"Ran" means the function executed on a synthesized input and returned a value. It is not a reproduction of the paper's results. This record, dated, as JSON: /api/ran/2306.00977.json · badge: /api/ran/2306.00977.svg

Each repository's role is a label: a classifier's output recorded on the link, or, where no role is recorded, what the Papers with Code archive marked. The line under it names its basis. Neither says who wrote the code.

RepositoryRoleRan
ywyue/agile3d community (classifier label)
classifier basis not recorded on the link
7 ran · 0 no recorded run
FunctionStatusWhere it lives
PositionalEncoding1D Ran
ran on a synthesized input
ywyue/agile3d/models/position_embedding.py
code served (permissive licence) · get_code("7cd78dccb309f0aa")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API) and returned Tensor, ndim 2, dtype family float, shape (4, 8) as a later run recorded it. The image it ran in was not recorded. If your copy returns the same type, ndim and shape on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import torch
import math

torch.manual_seed(0)

# PositionalEncoding1D takes two parameters:
# - d_model: dimension of the model (must be even)
# - length: length of positions

ARGS = {
    "d_model": 8,
    "length": 4
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
torch — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type Tensor, ndim 2, dtype family float, shape (4, 8) — shape (4, 8) as a later run recorded it
dtype family
int means an integer dtype; float means any other dtype the instrument saw, bool and complex included (fixture_check.py records only these two families)
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block
build_mask_criterion Ran
ran on a synthesized input
ywyue/agile3d/models/criterion.py
code served (permissive licence) · get_code("47cf8934a7785c31")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API) and returned SetCriterion. The image it ran in was not recorded. If your copy returns the same type on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import torch
import torch.nn as nn
from dataclasses import dataclass

torch.manual_seed(0)

@dataclass
class Args:
    bce_loss_coef: float
    dice_loss_coef: float
    losses: list
    aux: bool
    num_decoders: int
    hlevels: list

ARGS = {
    'args': Args(
        bce_loss_coef=1.0,
        dice_loss_coef=1.0,
        losses=['bce', 'dice'],
        aux=False,
        num_decoders=2,
        hlevels=[0, 1]
    )
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
torch — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type SetCriterion — a SetCriterion output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a SetCriterion output the recorded ndim 0 and dtype family float describe that coercion, not the output, and are not stated
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block
compute_averages Ran
ran on a synthesized input
ywyue/agile3d/compute_ap.py
code served (permissive licence) · get_code("041a1bcf15d34115")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API) and returned dict. The image it ran in was not recorded. If your copy returns the same type on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import numpy as np

# The function compute_averages takes one parameter: aps
# Looking at the function body, aps is indexed as aps[d_inf, :, indices]
# where d_inf = 0 (first dimension), : means all classes (second dimension),
# and indices are into opt_overlaps (third dimension)

# opt_overlaps has shape (10,) based on: np.append(np.arange(0.5, 0.95, 0.05), 0.25)
# which creates [0.5, 0.55, 0.6, 0.65, 0.7, 0.75, 0.8, 0.85, 0.9, 0.25]

# The function iterates over li in range(1) for ['agnostic'] (1 class)
# It accesses aps[d_inf, li, ...] where li goes from 0 to 0
# So aps needs shape (1, num_classes, num_overlaps) at minimum
# We'll use (1, 1, 10) to match the 1 class in the loop and 10 overlaps

# Create aps with shape (1, 1, 10) containing valid AP values
aps = np.random.rand(1, 1, 10)

ARGS = {
    'aps': aps
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
numpy — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type dict — a dict output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a dict output the recorded ndim 0 and dtype family float describe that coercion, not the output, and are not stated
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block
evaluate_matches Ran
ran on a synthesized input
ywyue/agile3d/compute_ap.py
code served (permissive licence) · get_code("4f5afa0c0f45149e")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API) and returned ndarray, ndim 3, dtype family float, shape (1, 1, 10) as a later run recorded it. The image it ran in was not recorded. If your copy returns the same type, ndim and shape on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import numpy as np
import tempfile
import os

# Create a temporary results file with test data
temp_dir = tempfile.mkdtemp()
results_file = os.path.join(temp_dir, "results.txt")

# Write sample results data
# Format: space-separated values where column 3 is clicks_num and column 4 is iou
with open(results_file, 'w') as f:
    # Write lines for clicks_num=1 with various IoU values
    f.write("image1 0 0 1 0.85\n")
    f.write("image2 0 0 1 0.72\n")
    f.write("image3 0 0 1 0.91\n")
    f.write("image4 0 0 1 0.55\n")
    f.write("image5 0 0 1 0.68\n")
    # Write some lines with different clicks_num to be filtered out
    f.write("image6 0 0 2 0.95\n")
    f.write("image7 0 0 2 0.80\n")

ARGS = {
    "results_file": results_file,
    "clicks_num": 1,
    "len_gt_instances": 5,
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
numpy — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type ndarray, ndim 3, dtype family float, shape (1, 1, 10) — shape (1, 1, 10) as a later run recorded it
dtype family
int means an integer dtype; float means any other dtype the instrument saw, bool and complex included (fixture_check.py records only these two families)
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block
get_num_instances Ran
ran on a synthesized input
ywyue/agile3d/compute_ap.py
code served (permissive licence) · get_code("0fb3d34bf75e3afd")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API) and returned int, ndim 0, dtype family int, shape () as a later run recorded it. The image it ran in was not recorded. If your copy returns the same type, ndim and shape on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import numpy as np
import pandas as pd
import tempfile
import os

# Create a temporary CSV file with the expected format
temp_dir = tempfile.gettempdir()
result_file = os.path.join(temp_dir, "test_result.csv")

# Create sample data with the columns: id, scene_id, instance_id, num_clicks, iou
data = {
    'id': [1, 2, 3, 4, 5, 6],
    'scene_id': [1, 1, 1, 2, 2, 2],
    'instance_id': [10, 11, 10, 20, 21, 20],
    'num_clicks': [1, 2, 1, 3, 2, 1],
    'iou': [0.85, 0.90, 0.82, 0.88, 0.91, 0.87]
}

df = pd.DataFrame(data)
# Write with space separator as expected by the function
df.to_csv(result_file, sep=' ', index=False, header=False)

ARGS = {
    'result_file': result_file
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
numpy, pandas — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type int, ndim 0, dtype family int, shape () — shape () as a later run recorded it
dtype family
int means an integer dtype; float means any other dtype the instrument saw, bool and complex included (fixture_check.py records only these two families)
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block
make_scan_transforms Ran
ran on a synthesized input
ywyue/agile3d/datasets/InterMultiObj3DSegDataset.py
code served (permissive licence) · get_code("716015228c6cbc1e")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API) and returned bool, ndim 0, dtype family float, shape () as a later run recorded it. The image it ran in was not recorded. If your copy returns the same type, ndim and shape on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import torch
import numpy as np

torch.manual_seed(0)

ARGS = {
    'split': 'train'
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
numpy, torch — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type bool, ndim 0, dtype family float, shape () — shape () as a later run recorded it
dtype family
int means an integer dtype; float means any other dtype the instrument saw, bool and complex included (fixture_check.py records only these two families)
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block
shift_scale_points Ran
ran; the contract check could not run (RAISES, DEP_MISSING or TIMEOUT)
ywyue/agile3d/models/position_embedding.py
code served (permissive licence) · get_code("818099bdffc93897")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1) and returned Tensor, ndim 3, dtype family float, shape (2, 4, 3) as a later run recorded it. The image it ran in was not recorded. If your copy returns the same type, ndim and shape on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import torch

torch.manual_seed(0)

# Create pred_xyz: B x N x 3
# B=2 (batch), N=4 (num points), 3 (xyz coordinates)
pred_xyz = torch.randn(2, 4, 3)

# Create src_range: [[B x 3], [B x 3]] - min and max XYZ coords
# B=2, 3 coordinates
src_min = torch.tensor([[0.0, 0.0, 0.0], [1.0, 1.0, 1.0]])
src_max = torch.tensor([[10.0, 10.0, 10.0], [11.0, 11.0, 11.0]])
src_range = [src_min, src_max]

# Create dst_range: [[B x 3], [B x 3]] - min and max XYZ coords
# B=2, 3 coordinates
dst_min = torch.tensor([[-1.0, -1.0, -1.0], [0.0, 0.0, 0.0]])
dst_max = torch.tensor([[1.0, 1.0, 1.0], [2.0, 2.0, 2.0]])
dst_range = [dst_min, dst_max]

ARGS = {
    "pred_xyz": pred_xyz,
    "src_range": src_range,
    "dst_range": dst_range,
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1
Imports
torch — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type Tensor, ndim 3, dtype family float, shape (2, 4, 3) — shape (2, 4, 3) as a later run recorded it
dtype family
int means an integer dtype; float means any other dtype the instrument saw, bool and complex included (fixture_check.py records only these two families)
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block

Also present in other papers' code

Of 7 samples lifted from this paper's linked repositories, 7 carry a function identity and 0 do not. Of the 7 distinct functions with an identity, 1 are also present in other papers whose code we hold (1 with at least one paper dated earlier, 0 the same month, 1 later) and 6 are only here among the 52,977 papers whose code Syntology holds (as of 2026-09-26).

Only here means: among the 52,977 papers whose code Syntology holds (51,887 with lifted samples, 1,090 through their assemblies only), as of 2026-09-26, no other paper's lifted samples or current assembly bricks contain an identical function body after normalisation.

FunctionAlso present inEarliest-datedOthers named
shift_scale_points 11 other papers
3 dated earlier, 8 later · 1 only as an assembly brick
2109.08141 (2021-09)
An End-to-End Transformer Model for 3D Object Detection
2210.03105 (2022-10)
Mask3D: Mask Transformer for 3D Semantic Instance Segmentation
2301.02508 (2023-01)
End-to-End 3D Dense Captioning with Vote2Cap-DETR
and 8 more not named here

Presence, not direction: an identical function body in two papers' repositories does not say who wrote it, which came from which, or that either paper's method rests on it. Matching is exact after normalisation, so a renamed or edited function is not matched. Dates are when papers appeared, not when code was written. Only .py files the harvester lifted are covered; notebooks are not read. A new-style arXiv id YYMM.NNNN dates a paper to the month 20YY-MM; otherwise Paper.year dates it to the year; otherwise it is undated. Dates are when the paper appeared, not when the code was written. Read 2026-11-22.

Repositories linked to this paper

Some links come from the archived Papers with Code dataset (CC BY-SA 4.0): attribution and licence.

Abstract

During interactive segmentation, a model and a user work together to delineate objects of interest in a 3D point cloud. In an iterative process, the model assigns each data point to an object (or the background), while the user corrects errors in the resulting segmentation and feeds them back into the model. The current best practice formulates the problem as binary classification and segments objects one at a time. The model expects the user to provide positive clicks to indicate regions wrongly assigned to the background and negative clicks on regions wrongly assigned to the object. Sequentially visiting objects is wasteful since it disregards synergies between objects: a positive click for a given object can, by definition, serve as a negative click for nearby objects. Moreover, a direct competition between adjacent objects can speed up the identification of their common boundary. We introduce AGILE3D, an efficient, attention-based model that (1) supports simultaneous segmentation of multiple 3D objects, (2) yields more accurate segmentation masks with fewer user clicks, and (3) offers faster inference. Our core idea is to encode user clicks as spatial-temporal queries and enable explicit interactions between click queries as well as between them and the 3D scene through a click attention module. Every time new clicks are added, we only need to run a lightweight decoder that produces updated segmentation masks. In experiments with four different 3D point cloud datasets, AGILE3D sets a new state-of-the-art. Moreover, we also verify its practicality in real-world setups with real user studies. Project page: https://ywyue.github.io/AGILE3D.

For agents

The same record, over MCP at https://syntology.ai/mcp:

get_harvested_code_for_paper("2306.00977")
get_code_for_paper("2306.00977")
have("2306.00977")

The run record, dated, one paper per request, free:

curl https://syntology.ai/api/ran/2306.00977.json

A badge for a README (the split and the date, never a ratio):

[![Syntology run record](https://syntology.ai/api/ran/2306.00977.svg)](https://syntology.ai/paper/2306.00977)

Connect an agent — have() is free.